Nutrient Pulse Reactor (DIY DyMiCo filter)

[dartier]

This is the comparison of the probes. I finally got both on the same scale so it is easier to see. There is quiet a bit of difference, especially near the zero to higher mV ranges. The curves look different too. This may be a result of the way Apex is visually rounding the data.

I still don't have an obvious nitrate knee, at least visually to me, even with a second dose of carbon.

Aaron

The first 3 cycles have a knee that is present. The last 2, not that I can see. The easy way to see it, other than waiting for Klaus to invert the graph and apply slope indicators , is to hold a business card up against the monitor. Line up the straight edge of the card to measure the slope. In the first 3 graphs you will see that a knee occurs and after that point the slope becomes more vertical until the low point (high point) is reached and the cycle is terminated.

Dennis

 

[OzIA]

Yeah but look at the other probe (lighter color). it actually bends the other direction. Is the knee you are seeing just the Apex visually smoothing the graph?

 

[OzIA]

I can't remember. Is the assumption that the knee will occur at different points during the cycle, based on the bio-load?

If we are thinking it should occur at the same value each time, why not just run a few test cycles to determine that point, say every week or so and then adjust the set points in the algorithm to that each week?

 

[dartier]

Yeah but look at the other probe (lighter color). it actually bends the other direction. Is the knee you are seeing just the Apex visually smoothing the graph?

Oh, I see what you mean. That is not good. Are the probes close together? The lighter coloured one seems to be reading completely differently.

Dennis

 

[dartier]

I can't remember. Is the assumption that the knee will occur at different points during the cycle, based on the bio-load?

If we are thinking it should occur at the same value each time, why not just run a few test cycles to determine that point, say every week or so and then adjust the set points in the algorithm to that each week?

The time interval before the knee occurs will be dependent on.

1. The amount of fresh water added to the filter.
2. The amount of nitrate present in the water added.
3. The size of the biofilm that is denitrifying the incoming water.

So it is not really something that can be predicted exactly. However the fact that you have managed to get your nitrate down so low, using only fixed time period cycling probably shows that there is a fair amount of leeway to get a positive result.

Until we do more testing with different hardware that provides more detailed data points, it will be difficult to know for sure if we are on the right path by thinking we can reliably detect the knee. I am still optimistic that it can be done.

Dennis

 

[OzIA]

Ok, so the time interval varies, I can see that but If you perform some sort of test, maybe weekly, to determine at what ORP level it occurs, can you then trigger based on an ORP value and then flush or add carbon, based on that set point?

 

[OzIA]

The probes are about an inch apart. One is newer than the other.

Also, I believe I have found a source of some of my issues. These BRS dosers suck. Funny, auto correct wants to call them "losers". It mush be smarter than i imagined.

Anyway, I tried three of them, all have dead spots. Some are worse than the others. I measured one and it delivered 9.5 ml in 5 minutes. That is certainly not the level of precision we need and they seem to imply, with their 1.1 ml / min claim.

Aaron

 

[dartier]

Ok, so the time interval varies, I can see that but If you perform some sort of test, maybe weekly, to determine at what ORP level it occurs, can you then trigger based on an ORP value and then flush or add carbon, based on that set point?

One of the issues is that ORP is not deterministic, or at least for our purpose it will seem that it is not. Meaning that the level is not exact and will be responding to the varying ratios of compounds within the water column, most of which we don't care about anyway. So it would be difficult to say with certainty that when we reach ORP xx, the nitrate has been depleted. But if we are able to use the nitrate knee point, then the ORP level really becomes irrelevant and most ORP probes should work, as long as they are at least linear and functional.

However, the DyMiCo folks appear to use a lower bound for ORP and key on it. They also require the ORP probe to be calibrated monthly, so they are at least attempting to have an idea that -250mv is really -250mv based on a calibration against a standard.

So depending on how successful we are at being able to pinpoint the occurrence of a knee, we may also have to implement similar logic as a worst case scenario.

Dennis

 

[dartier]

This is the comparison of the probes. I finally got both on the same scale so it is easier to see. There is quiet a bit of difference, especially near the zero to higher mV ranges. The curves look different too. This may be a result of the way Apex is visually rounding the data.

I still don't have an obvious nitrate knee, at least visually to me, even with a second dose of carbon.

Aaron

One thing that crossed my mind, is that you are dosing C on every flush, and the DyMiCo manual hints that they dose C once an hour. So I am wondering if the look of the chart changed because of the additional C dose that you added. For the probe that appears to show a knee, the last cycle that showed a knee was the one where you added the extra dose. The cycles after that first cycle, do not appear to show a knee.

At this point, you may only be cycling once a hour anyway, so you may be dosing in a similar fashion to a DyMiCo filter that is still cycling at the 1 hour minimum. I would remove the 2nd C dose and see if your graph returns to the previous pattern.

Dennis

 

[OzIA]

Latest. Old probe ORpHX in Green (scale on right) New probe (pH) scale on left.

There is a lot of variance. I think the old one is dying.

 

[dartier]

I think you are right. The old probe's waveform looks unusual, whereas the new probe's waveform looks like what we would expect to see. What brand of probes are these by the way? Are they both Apex ORP probes?

 

[OzIA]

One is Apex (old) and the new is the BRS version.

 

[OzIA]

So these readings would represent 75 to -200 mV peak to peak. The avg would be -90 mV.

 

[OzIA]

Sorry, avg is -80 mv.

 

[dartier]

One is Apex (old) and the new is the BRS version.

I had a look at what BRS sells. Their in house double junction ORP probe, which is what i assume you have, looks pretty decent. I would retire the one that is returning varying readings and only reference the new probe in your control loop.

Dennis

 

[dartier]

Not so much an update on my filter, but an update on my tank that the filter is waiting on. I stopped my Miracles to check on the progress and it was a good thing that I did. They had forgot to build my stand, so the tank delivery is delayed waiting for the stand to be powder coated this week. I am taking the opportunity to have them install some cross braces as the length of tank unbraced tank is making me nervous. Well the length and other forum members!

Here is a photo of the tank at Miracles.

The rest of the photos are in my build thread at:

http://www.reefcentral.com/forums/showthread.php?t=2607623

Dennis

 

[OzIA]

No two ways about it, I need to find some way to calibrate these ORP probes.

The old probe finally straightened out and started looking like the newer BRS model over the holiday. However, there is still a big difference between the two.

Since I am using the pH channel for ORP, I had to calibrate them using a pH probe. I used the same probe, hoping to at least eliminate one variable but they still ran about 100 mV different.

The other day I switched where each was plugged into the Apex. The variation immediately followed the probe.

 

[dartier]

No two ways about it, I need to find some way to calibrate these ORP probes.

The old probe finally straightened out and started looking like the newer BRS model over the holiday. However, there is still a big difference between the two.

Since I am using the pH channel for ORP, I had to calibrate them using a pH probe. I used the same probe, hoping to at least eliminate one variable but they still ran about 100 mV different.

The other day I switched where each was plugged into the Apex. The variation immediately followed the probe.

I don't know about Apex, but with the Reef Angel Orp module, you get a BNC terminator that you swap in for calibration. The terminator is equal to 0 mv. That along with a 400 mv calibration packet is what I will be using to calibrate my ORP probe. Does Apex offer something similar?

Dennis

 

[OzIA]

Not that I know of but I am not sure it matters. Remember, I have to use the channel as a pH because Apex won't read negative ORP values. I got in an an argument with Apex about the issue. It seems like a simple software issue to me. If the hardware is capable of reading - mV values as a pH probe, then it seems like a simple fix to allow the ORP values to be read at a negative value but the tech at Apex said "it wasn't that simple".

There is an ORP calibration screen but it doesn't seem to do anything. Maybe because I have not probes set up as ORP, only pH.

 

[dartier]

Not that I know of but I am not sure it matters. Remember, I have to use the channel as a pH because Apex won't read negative ORP values. I got in an an argument with Apex about the issue. It seems like a simple software issue to me. If the hardware is capable of reading - mV values as a pH probe, then it seems like a simple fix to allow the ORP values to be read at a negative value but the tech at Apex said "it wasn't that simple".

There is an ORP calibration screen but it doesn't seem to do anything. Maybe because I have not probes set up as ORP, only pH.

Sounds like a software limitation to me as well. Are the PH modules and ORP modules the same with only the probe being different (on Apex)? If so why not calibrate your ORP probes using the PH calibration? You would need a BNC terminator to give you 0 mv and an ORP standard (400 mv?). That way you could calibrate your probes even though you have them setup as a PH probes. I would be curious to see if this idea works.

Dennis